Exploring the Thalamus

Exploring the Thalamus

First Edition

S. Murray Sherman

Department of Neurobiology State, University of New York, Stony Brook, New York

R.W. Guillery

Department of Anatomy, University of Wisconsin School of Medicine, Madison, Wisconsin

ACADEMIC PRESS

An Imprint of Elsevier

San Diego San Francisco New York Boston London Sydney Tokyo

Table of Contents

Cover image

Title page

Copyright page

Preface

Abbreviations

Chapter I: Introduction

A Thalamic Functions: What Is the Thalamus and What Is It For?

B The Thalamus as a Part of the Diencephalon: The Dorsal Thalamus and the Ventral Thalamus

C The Overall Plan of the Next Nine Chapters

Chapter II: The Nerve Cells of the Thalamus

A On Classifying Relay Cells

B Interneurons

C The Cells of the Thalamic Reticular Nucleus

D Summary

E Some Unresolved Questions

Chapter III: The Afferent Axons to the Thalamus

A A Functional Classification of Afferents to the Thalamic Nuclei

B Afferent Axon Types as Seen Light Microscopically

C Electron Microscopic Appearance of the Afferent Axon Terminals and Their Synaptic Relationships

D Afferents from Interneurons and Reticular Cells

E GABA Immunoreactive Afferents

F Afferents to the Thalamic Reticular Nucleus

G Some Problems of Synaptic Connectivity Patterns

H Summary

I Some Unresolved Questions

Chapter IV: Intrinsic Cell Properties

A Cable Properties

B Membrane Conductances

C Summary and Conclusions

D Some Unresolved Questions

Chapter V: Synaptic Properties

A Ionotropic and Metabotropic Receptors

B Synaptic Inputs to Relay Cells

C Inputs to Interneurons and Reticular Cells

D Summary

E Some Unresolved Questions

Chapter VI: Function of Burst and Tonic Response Modes in the Thalamocortical Relay

A Rhythmic Bursting

B Effect of Response Mode on Transmission from Relay Cells to Cortical Cells

C Control of Response Mode

D Summary

E Some Unresolved Questions

Chapter VII: Maps in the Brain

A Introduction

B Early Arguments for Maps

C Clinical and Experimental Evidence for Maps in the Geniculocortical Pathway

D Multiple Maps in the Thalamocortical Pathways

E Abnormal Maps in the Visual Pathways

F Maps in the Thalamic Reticular Nucleus

G Summary

H Some Unresolved Questions

Chapter VIII: Two Types of Thalamic Relay

A The Basic Categorization of Relays

B The Evidence That There Are Two Distinct Types of Afferent That Go from Neocortex to Some Thalamic Nuclei

C The Relationship of First and Higher Order Relays to the Thalamic Reticular Nucleus

D Summary

E Some Unresolved Questions

Chapter IX: Drivers and Modulators

A Drivers and Modulators in the Lateral Geniculate Nucleus

B The Geniculate Input to Cortex as a Driver

C Tonic and Burst Modes in Thalamic Relay Cells

D Key Differences between Drivers and Modulators

E The Sleeping Thalamus

F Can Extradiencephalic GABAergic Inputs to Thalamus Be Drivers?

G Summary

H Some Unresolved Questions

Chapter X: Overview

Bibliography

Index

Copyright

The following items were taken with permission from previously published sources:

Chapter II: Fig. 1 from Kaas et al., © 1972 Karger, Basel. Fig. 3 from Rodieck and Brening, © 1983 Karger, Basel, and based on data of Boycott and Wässle (1974). Fig. 6A,B from Bartlett and Smith (1999). Fig. 6C,D from Winer and Morest, J. Comp. Neurol. 221, 1–30, © 1983 John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Fig. 6E,F from Guillery, J. Comp. Neurol. 128, 21–50, © 1966 John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Chapter III: Fig. 1 from Lachica and Casagrande (1988). Reprinted with the permission of Cambridge University Press. Fig. 2 from Pallas et al., J. Comp. Neur. 349, 343–362, © 1994 John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Fig. 3 from Sur et al. (1987). Fig. 4 from Guillery, J. Comp. Neurol. 128, 21–50, © 1966 John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Fig. 10 from Cox et al., J. Comp. Neurol. 366, 416–430, © 1966 John Wiley & Sons. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Fig. 11 from Uhlrich et al. (1991). Chapter VII: Fig. 1A,B based on Cajal (1995), Histology of the Nervous System of Man and Vertebrates (translated by N. and L. W. Swanson), Oxford University Press, Footnote 2 from Walls (1953), University of California Publications in Physiology Vol. 9. Chapter IX: Fig. 1A from Fig. 3A of Mastronarde (1997). Fig. 1B redrawn from Fig. 2A of Tsumoto et al. (1978). Exp. Brain Res. 32, 345–364, © 1978 Springer-Verlag. Data in Figs. 1C,D kindly provided from the authors for replotting: Fig. 1C redrawn from Usrey et al. (1998). Nature 395, 384–386. Fig. 1D redrawn from Reid and Alonso (1996). Curr. Opin. Neurobiol. 6, 475–480, © 1996, with permission from Elsevier Science.

Copyright © 2001 by S. Murray Sherman and R. W Guillery

All Rights Reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Permissions may be sought directly from Elsevier’s Science and Technology Rights Department in Oxford, UK. Phone: (44) 1865 843830. Fax: (44) 1865 853333, e-mail; permissions@elsevier.co.uk. You may also complete your request on-line via the Elsevier homepage: http://www.elsevier.com by selecting Customer Support and then Obtaining Permissions.

Academic Press

An Imprint of Elsevier

525 B Street, Suite 1900, San Diego, California 92101-4495, USA

http://www.academicpress.com

Academic Press

A Harcourt Science and Technology Company

http://www.academicpress.com

Library of Congress Catalog Card Number: 00-104281

ISBN-13: 978-0-12-305460-9

ISBN-10: 0-12-305460-5

PRINTED IN THE UNITED STATES OF AMERICA

05 QW 9 8 7 6 5 4 3 2

Preface

S. Murray Sherman; R.W. Guillery

The title of this book, Exploring the Thalamus, is intended to convey the sense that there is a great deal still unknown about the thalamus and that our aim in writing has been to show some of the major and minor roads, the footpaths, and the frank wildernesses that still need to be explored. The thalamus, in terms of its detailed connectivity patterns, the functioning of its circuitry, and, perhaps most interestingly, its functional relationship to the cerebral cortex, is largely terra incognita. The cortex depends critically on the messages it receives from the thalamus. It receives very little else. Understanding cortical functioning will depend on understanding the thalamic inputs that are the necessary first step in cortical processing. There is a serious sense in which one can regard the thalamus as the deepest layer of the cortex. That is, cortex and thalamus depend very closely on each other; neither would amount to much without the other, and if we are to understand the workings of either, we must, as we try to show in what follows, be able to understand the messages that each is sending to the other. Our main focus on the thalamus in this book will be on its relation to cortical function. Perhaps this focus on thalamocortical interrelationships should have been a part of our title, but we wanted a brief title and we had in mind a book on the thalamus, not one on the cortex. We are looking at ways to explore how thalamic organization should influence our views of cortical function; we are not looking to provide a general account of the thalamus as an entity in itself.

In a preface, authors are expected to say for whom the book has been written. We hope that this book will serve to introduce graduate students, postdoctoral fellows, and investigators who need to learn about the thalamus to some of the interesting aspects of the subject. Also, since many view the thalamus as an uninteresting, mechanical relay of peripheral messages to cortex that is already well understood, we have tried to explain that this view is far too simplistic and that there are many important problems about the function and the structure of thalamus that remain unrecognized and unresolved. One of our aims has been to persuade colleagues that these problems are of interest and worth significant research investment. We have tried to make each chapter more or less independent, so that perhaps one or another can be assigned as course reading for graduate students. This entails some repetition from one chapter to another; we trust that it is not excessive for those who are motivated to read the whole thing.

Our thoughts about who would read the book were not our main focus when we first started discussions and rough drafts almost 10 years ago. Rather, we undertook the task initially because we found that the thoughts, the discussions, and the arguments that have accompanied the job of writing were sufficient stimulus in themselves. As we wrote, exchanged drafts, sent each chapter back and forth many times to be annotated, corrected, reannotated, and recorrected, we gradually learned a great deal about our subject (and about each other). The real truth is that we wrote the book for ourselves, and once we had that aspect of the writing fairly in hand, we worked hard to make it accessible to others.

We have not attempted to present a complete and coherent view of everything that is known about the thalamus. We have instead followed arguments and lines of inquiry that can lead to new questions, interesting thoughts, or new experimental approaches. Knowledge of the thalamus is extraordinarily patchy. The thalamus is divided into many different nuclei. There are some thalamic nuclei that have been studied in considerable detail, and others about which we know almost nothing. Our plan in writing the book has been to assume that there is a basic ground plan for the thalamus. Although there are often important differences between one thalamic nucleus and another, in one species or between species, there is yet a common pattern of organization seen over and again in essentially all thalamic nuclei. We have tried to explore the nature of this common pattern and to ask questions about its functional significance.

We have both spent the greater part of our careers studying the visual pathways, and it won’t take a very subtle reading of the book to recognize this. We turn to the visual relay in the thalamus repeatedly not only because this is the part we know best but also because in our readings and in our discussions with colleagues we find that the visual relay has, time and again, received more detailed experimental study than other thalamic relays. The visual relay may well have some special characteristics that distinguish it from other relays. In some instances this is clear, and we recognize it. However, in many instances it is reasonable to treat the visual relay as an exemplar of thalamic relays in general, and in many parts of the book, that is how we have approached the analysis of thalamic functions. This approach raises important questions about non-visual parts of the thalamus, and our expectation is that the comparisons will stimulate further study of these questions.

We have stressed that this book is not a complete inventory of all that is known about the thalamus. There are many important references we have not cited, and there are several lines of inquiry that we have not included. We say virtually nothing about the development or the comparative anatomy of the thalamus even though each is an extremely interesting subject in its own right. They should perhaps form the nucleus of another book. Nor do we cover the clinical aspects of thalamic dysfunction, another potentially interesting area, although it seems likely to us that this will become of greater interest once we know more about some of the basic ground rules of thalamic function and connectivity that are still missing from our current knowledge. For instance, the thalamus has long been implicated in epilepsy and certain sleep disorders, it is related to the production of pathological pain, and there is new interest in the thalamus as a particularly interesting site of pathology in schizophrenia and other cognitive problems. The complexity of the two-way links between thalamus and cortex and the limited nature of our knowledge about these links, especially in the human brain, make interpretations of clinical conditions extremely difficult and often rather tenuous, and we have not addressed them in this book.

We have tried to achieve two major aims in the book. The first is to look at many of the outstanding puzzles and unanswered questions that arise as one studies the structural and functional organization of the thalamus. The second aim, growing out of a small proportion of these questions, is to move toward an understanding of the possible role(s) of the thalamus in cortical functions, so that some coherent suggestions about this role could be presented as the book proceeds. The first aim is summarized to a limited extent by a short list of Some Unresolved Questions that appears at the end of each chapter. These are not questions to which a student can find answers in the text. They are, rather, designed to focus on some of the issues that need to be resolved if we are to advance our understanding of the thalamus. They do not represent an exhaustive list, and the interested reader is likely to find a number of other questions that are currently unanswered and often unasked. The listed questions should be seen as representing a state of mind, and they are an important part of the book as a whole. They should lead to more questions, and they should point to paths that have perhaps never been explored or along which our predecessors have been lost in the past. We hope that by stimulating a questioning attitude to thalamic organization we will encourage a view of the thalamus as far more mysterious than is commonly taught. This clearly implies that our second aim, to understand the role of the thalamus, which we present in detail in the later chapters, can at best be only partially achieved. We present a view of the thalamus that is based on the classical view of it as a relay of ascending messages to cortex. However, we see it as a continually active relay, serving sometimes as a lookout for significant new inputs and at other times as an accurate relay that allows detailed analysis of input content in the cortex. This is based on the recognition of two distinct types of input to the thalamus, the drivers that carry the message and the modulators that determine how the message is transmitted to cortex. The former can carry ascending messages from the periphery as well as descending messages from cortex itself. These messages are generally mapped, giving them a definite locus in the environment or in some other part of the brain. In contrast, the latter either can be mapped and thus act locally like the drivers or can lack a mapped organization and then act globally. Recognizing that drivers can take origin in the cortex leads to an interesting new view of corticocortical communication because it stresses that messages that pass from one cortical area to another may be under the same set of modulatory controls in the thalamus as are the inputs that are passed to the cortex from the peripheral senses.

It is probable that many of the ideas we present in this book will prove wrong. Whether they are right or wrong, we have tried to make them stimulating. To quote Kuhn (1963) quoting Francis Bacon, Truth emerges more readily from error than from confusion, which provides our best justification for writing this book about a subject that in terms of the currently available literature is often extremely confusing.

Finally, both authors owe thanks and the book itself owes its existence to many people and organizations. A number of colleagues read an early draft of the book, and the comments and critical points that they raised have helped us to reorganize and correct a great deal of the book in terms of style, order of presentation, and content. We thank Paul Adams, Joe Fetcho, Sherry Feig, Lew Haberly, Carsten Hohnke, Jon Levitt, John Mitrofanis, and Phil Smith for their helpful comments. We recognize the amount of time and effort that they have contributed; we are most grateful for it and for the significant improvements that their careful readings have produced. The final version of this book is, of course, entirely our responsibility. All of our colleagues will likely find many places where they can write further instructive comments in the margins, and perhaps some of these will lead to useful explorations of the thalamus in the future. Majorie Sherman helped with the proof reading. Sherry Feig helped one of us (R.W.G.) learn how to draw on a computer. Both authors received support from the NIH while this book was being written (Grants EY03038, EY11409, and EY11494), and at the early stages R.W.G., while in the Department of Human Anatomy at Oxford, was supported by the Wellcome Trust. The initial stimulus for planning the book came from the year S.M.S. spent as a Newton-Abraham Visiting Professor at Oxford in 1985–1986.

Abbreviations

We have, as far as possible, avoided the use of abbreviations, except for a few that are commonly used and widely recognized for complex names. They are the following: AMPA, (R,S)-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; EPSP, excitatory postsynaptic potential; GABA, γ-aminobutyric acid; GAD, glutamic acid decarboxylase; IPSP, inhibitory postsynaptic potential; NMDA, N-methyl-d-aspartate.

Chapter I

Introduction

A Thalamic Functions: What Is the Thalamus and What Is It For?

The thalamus is the major relay to the cerebral cortex. It has been described as the gateway to the cortex. Almost everything that we can know about the outside world, or about ourselves, is based on messages that have to pass through the thalamus. It forms a relatively small cell group on each side of the third ventricle and can be seen most readily in a midsagittal section of the brain (Figure 1). It can be divided into several distinct nuclei or nuclear groups, each concerned with transmitting a characteristic type of afferent signal (visual, auditory, somatosensory, cerebellar, etc.) to a cytoarchitectonically distinct and functionally corresponding area or group of areas of the cerebral cortex (Figures 2 and 3) on the same side of the brain. This is the view of the thalamus that was developed during the 70 (plus) years up to circa 1950. It has served us well and is still the view presented in most textbooks. It was based on clinical observations related to postmortem study of the brain and on relatively crude experimental neuroanatomical methods: the Marchi method for staining degenerating myelin in pathways that had been cut or injured, and the Nissl method, which shows nerve cells grouped into functional nuclei or undergoing retrograde degeneration after their axons have been cut. These methods give results in terms of large populations of cells or axons and large areas of thalamus or cortex. Perhaps it was fortunate that modern methods for studying the detailed connectivity patterns of single cells or small groups of cells were not available when the thalamic connections were first being defined. If they had, it is probable that no one would have been able to see the larger thalamic forest for the details of the connectional trees. We shall start with the forest.

Figure 1 Midsagittal view of the cerebral hemispheres of a human, a monkey, a cat, and a rat (in size order) to show the position and relative size of the thalamus, which is indicated by diagonal hatching.

Figure 2 Schematic view of five sections through the thalamus of a monkey. The sections are numbered 1 through 5 and were cut in the coronal planes indicated by the arrows in the upper right midsagittal view of the monkey brain from Figure 1. The major thalamic nuclei in one hemisphere for a generalized primate are shown. The nuclei that are outlined by a heavier line and filled by diagonal hatching are described as first order nuclei (see text), and the major functional connections of these, in terms of their afferent (input) and efferent (output) pathways to cortex, are indicated in Figure 3. AD, anterodorsal nucleus; AM, anteromedial nucleus; AV, anteroventral nucleus; CM, center median nucleus; CN, caudate nucleus (not a part of the thalamus); H, habenular nucleus; IL, intralaminar (and midline) nuclei; LD, lateral dorsal nucleus; LGN, lateral geniculate nucleus; LP, lateral posterior nucleus; MGN, medial geniculate nucleus; PO, posterior nucleus; PU, pulvinar; TRN, thalamic reticular nucleus; VA, ventral anterior nucleus; VL, ventral lateral nucleus; VPI, VPL, VPM, inferior, lateral, and the medial parts of the ventral posterior nucleus or nuclear group.

Figure 3 The upper part shows the nuclei illustrated in Figure 2 , and the lower part shows a lateral (left) and a medial view of the hemisphere in a monkey to indicate the functional connections of the major first order thalamic nuclei. AUD, auditory; CING/AFF, cingulate/affective; MOT, motor; SOMSENS, somatosensory; VIS, visual.

The schematic view of thalamocortical relationships, summarized in Walker’s great book (1938) or in Le Gros Clark’s earlier review (1932), provided a powerful approach to thalamic function. Even though it was heavily dependent on relatively gross methods, it showed us how to divide up the thalamus and how to relate each of the resulting major thalamic nuclei or nuclear groups to one or another part of the cerebral cortex (see Figures 2 and 3). Above all, it demonstrated the extent to which the functions of any one part of the cerebral cortex must be dependent on its thalamic inputs. It should be noted that the thalamus is connected to the so-called neocortex. Olfactory cortex and hippocampal cortex are not neocortex and do not receive comparable thalamic afferents. Olfactory afferents represent the only pathway of a sensory system that does not have to go through the thalamus before it can reach the cortex. The following account deals with the relationships between thalamus and neocortex only.

Figure 1 shows the thalamus in relation to the rest of the cerebral hemisphere and shows that the thalamus is small relative to the whole cerebral hemisphere in all mammals. There are a great many more cortical cells than there are thalamic cells even though the neocortex depends on the thalamus for its major inputs.¹ Each major neocortical area depends on a well-defined thalamic cell group, that is, on a thalamic nucleus or group of nuclei. In the evolutionary history of mammals, an increase in the size of any one part of cortex generally relates to a corresponding increase in the related thalamic nuclei. The functionally best-defined cortical areas (visual, auditory, motor, etc.) depend for their functional specialization on the nature of messages that pass to that cortical area from the thalamus. The visual cortex is visual because it receives visual messages from the retina through its thalamic relay, and the same is true of the other thalamic nuclei outlined in bold and hatched in Figure 2.

Figure 2 shows some of the major thalamic nuclei in a simplified, schematic form for a generalized primate. Details differ for each species, and a number of nuclei are not included in the figure because they play no role or only a minor role in the rest of this book. However, the general relationships shown apply to all mammals. Figure 3 shows how some of these major thalamic nuclei are linked to specific, functionally or structurally defined cortical areas. Further details on individual thalamic nuclei and their connections can be found in Berman and Jones (1982) and Jones (1985). Figures 2 and 3 show that for some, but by no means for all of the thalamic nuclei, we can define the dominant or functionally driving afferents. That is, Figure 3 shows that the lateral geniculate nucleus is visual and the medial geniculate nucleus is auditory. The ventral posterior nucleus² is somatosensory, which is to say that the ascending pathways concerned with tactile stimuli and with stimuli related to the position and movements of body parts (kinesthesis) go to this nucleus as do pathways concerned with pain and temperature. For reasons detailed in Chapter III we treat the afferents from the cerebellum, related to movement control, that go to the ventrolateral and ventral anterior cells groups as drivers and also the mamillothalamic tract that sends information to the anterior thalamic nuclei about ongoing activity in the hippocampal formation and in the midbrain. These represent the major, known ascending driver inputs to the thalamus. We shall treat these ascending afferents as the drivers of the thalamic relay cells that they innervate, meaning by this that they are the afferents that determine the receptive field properties of the relay cells. Other afferents, which we shall treat as modulators, can modify the way that the message is transmitted, but they are not responsible for the main qualitative nature of the message that is conveyed to the cortex by the thalamic relay cell. We shall argue that each thalamic nucleus has drivers and modulators and that identifying the drivers for many thalamic nuclei is likely to be a key to understanding their functions, which are currently poorly defined.

The afferents to the other main thalamic nuclei, shown with lighter outlines and no shading in Figure 2 and unlabeled in Figure 3, are less straightforward and will be considered later (Chapters III and VIII). We have presented evidence that these nuclei receive their major driving afferents from the cerebral cortex itself and thus serve as a relay on corticocortical pathways, not as relays of subcortical afferents to cortex (Guillery, 1995; Sherman and Guillery, 1996). That is, they contain higher order relays. We define first order relays as those concerned with sending messages to the cortex about what is happening in the subcortical parts of the brain and higher order relays as those that provide a transthalamic relay from one part of cortex to another. It is to be noted that in primates the nuclei that contain the higher order circuits form more than half of the thalamus. How these indirect corticocortical relays relate to direct corticocortical connections is a challenging question considered in Chapters VIII and IX.

Although for many of the thalamic nuclei we can show how they serve to connect different cortical areas to sensory surfaces of the body or to other parts of the brain, we cannot readily demonstrate what it is that the thalamus does for the messages that are passed from ascending pathways to the cerebral cortex. Why don’t the ascending pathways go straight to the cortex? This question was always present, but was brought into striking focus in the 1960s when electron microscopists showed the complexities of the synaptic relationships in the thalamus (Szentágothai, 1963; Colonnier and Guillery 1964; Peters and Palay, 1966). Only about 20% of the synapses in the major relay nuclei like the lateral geniculate were then seen to come from the major ascending pathways (Guillery, 1969b), and recent figures show less than half that percentage contact the relay cells that are responsible for passing the information on to the cortex (Van Horn et al., 2000). Complex synaptic formations involving serial synapses and connections from local or distant inhibitory cells characterize all of the thalamic nuclei (e.g., Ralston and Herman, 1969; Jones and Powell, 1969; Morest, 1975), and most thalamic nuclei, in accordance with their shared developmental origin, have more or less the same general organizational plan. Exceptions are considered in later chapters.

The complexity of thalamocortical pathways was further increased by the demonstration of the connections shown in Figure 4. Not only is there a massive input from the deeper layers of the cerebral cortex back to the thalamus, but there is a specialized cell group adjacent to the thalamus, the thalamic reticular nucleus, that receives excitatory branches from the corticothalamic and thalamocortical axons and sends inhibitory axons back to the thalamus (Jones, 1985). The functional role of these reticular connections and of the complex synaptic arrangements found in the thalamus represented (and still represents) a challenging puzzle, a challenge that was greatly increased in recent years by the discovery of diverse transmitters, voltage- and ligand-gated ion channels, and receptors that contribute to the synaptic organization in the thalamic relay (see Sherman and Guillery, 1996; and Chapters V and VI). The functional control of membrane conductances depends on a highly complex interplay of afferent activity and local conditions that will be considered in Chapter IV. These conditions in turn determine the way in which a thalamic cell responds to its inputs and thus determine how messages that come into the thalamus are passed on to cortex. This, the manner in which a thalamic cell passes messages on to cortex, is not constant but depends on the attentive state of the whole animal (waking or sleeping) and probably on the local salience of a particular stimulus or group of stimuli as well: are they new, threatening, interesting, or merely a continuation of prior conditions? (See Chapter VI.)

Figure 4 Schematic view of the interconnections between the first order thalamic nuclei, cerebral cortex, and the thalamic reticular nucleus (TRN) as seen on coronal sections of thalamus and cortex. The nuclei labeled A and B are connected with distinct sectors of the reticular nucleus also labeled A and B, and with distinct cortical areas. The connections are shown for components A and B only. The cortical pyramids that give off collateral branches to the reticular nucleus lie in cortical layer 6. CN, caudate nucleus.

When one considers the factors relevant to how the transfer of messages is controlled or gated in the thalamus, it is probable that more than one functionally significant mechanism will be apparent once we have a clear understanding of these aspects of thalamic organization. That is, there are likely to be several more or less distinct functional roles for the synaptic arrangements present in the thalamus. They may be active at different times, or they may have concurrent effects. Two mechanisms that have received significant attention in the recent past have roles in sleep and in the production of epileptic discharges (McCormick and Bal, 1997; Steriade et al., 1993). A third that is corning into focus currently and will be addressed later in the book (Chapters VI and IX) concerns how the role of the relays may change in relation to different behavioral states and may relate to attentional mechanisms. All three involve the circuit going through the thalamic reticular nucleus that was mentioned previously (Figure 4; see Jones, 1985). We anticipate that the role of first and higher order thalamic circuits in passing messages to the cortex will follow the same basic ground rules. That is, whatever it is that the thalamus does for the major ascending pathways, it is likely to be doing something very similar for corticocortical communication. Understanding what it is that the thalamus does should not only help us to understand the functional organization of sensory pathways in relation to perception but also throw new light on perceptual and cognitive functions that have in the past been largely or entirely ascribed to corticocortical interconnections (Zeki and Shipp, 1988; Felleman and van Essen, 1991; Salin and Bullier, 1995).

There is one interesting corollary to the above. If the thalamus acts to control the way that information is relayed to the cortex, then it may be a mistake to look for it to act as an integrator of distinctive inputs as well. At present, the most detailed information available on thalamic relays shows that information from the ascending pathways is being passed to cortex without significant change in content. That is, there are thalamic nuclei that receive afferents from more than one source, but currently there is no evidence that the two inputs in such nuclei interact to produce a significant change in the content of the input messages. The two pathways appear to run in parallel, with little or no interaction. The motor pathways that go through the ventral anterior/ventral lateral nuclei may prove an interesting exception (see Chapter III), but currently this appears to be an open question.

In this book we are concerned with exploring the way in which thalamic functions relate to cortical functions. Outputs of the thalamus that link it to other cerebral centers, particularly to the striatum and to the amygdala, represent a relatively small although important part of the thalamic relay. They play no role or only an indirect role in influencing neocortical activity, and for this reason we will not be exploring them further. We shall argue that there is likely to be a basic thalamic ground plan that represents the way in which the thalamus transmits messages from its input to its output channels. It seems probable that this ground plan will apply to all thalamic relays, and possibly, when we understand how the thalamus relates to the cortex, the nature of this thalamic relay will help us to understand the function of the currently even more mysterious pathways to other cerebral centers.

B The Thalamus as a Part of the Diencephalon: The Dorsal Thalamus and the Ventral Thalamus

The term thalamus is commonly used to apply to the largest part of the mammalian diencephalon, the dorsal thalamus, and we will generally use it in this sense in what follows. However, it is important to recognize that there are several subdivisions of the diencephalon and to look briefly at all of them before focusing on just two subdivisions: the large dorsal thalamus and the smaller, but closely related, ventral thalamus.

Figure 5 shows relationships in the diencephalon at a relatively early stage of development. The upper figure, which is a view of the lateral aspect of the brain early in development, shows that the most dorsal part of the diencephalon is the epithalamus, which in the adult is made up of the habenular nuclei, a few other small dorsally placed nuclei, and the related pineal body. These will not be of further concern to us, nor will two more ventral cell groups, the subthalamus, which is not shown in Figure 5 and is involved with motor pathways, and the hypothalamus, which plays a vital role in neuroendocrine and visceral functions. In this book we are concerned solely with the dorsal thalamus and with a major derivative of the ventral thalamus, the thalamic reticular nucleus.³ These two are closely connected by the two-way links shown in Figure 4, and it is reasonable to argue that neither can function adequately without the other. Figure 5 shows that originally, during development, the ventral thalamus lies ahead of (rostral to) the dorsal thalamus. The dotted lines in the schematic views of Figure 5 stress an important relationship between the dorsal and the ventral thalamus because they show that lines of communication between the dorsal thalamus and the telencephalon, which includes the cerebral cortex, must pass through the ventral thalamus. This is a key relationship and is maintained even when the ventral thalamic derivative, the thalamic reticular nucleus, moves into its adult position lateral to the dorsal thalamus as shown in Figures 2 and 3.

Figure 5 Schematic views of two sections through a 14-day postconception fetal mouse brain, based on photographs in Schambra et al. (1992) . The upper figure shows a parasagittal section in which the positions of the epithalamus, the dorsal thalamus, the ventral thalamus, and the hypothalamus within the diencephalon are shown (EPI, DORSAL, VENTRAL, HYPO). The lower figure shows a section cut transversely in the oblique plane, indicated by the arrow, that includes these four diencephalic parts and the optic chiasm (OX). The subthalamus is not included in these figures. The interrupted lines show the course of the fibers that link the dorsal thalamus to the telencephalon. LV, lateral ventricle; IIIV, third ventricle.

1 THE DORSAL THALAMUS

In most mammalian brains, and most strikingly in the primate brain, the dorsal thalamus is by far the largest part of the diencephalon. In its size and complexity it is closely related to the development of the cerebral cortex. It can be defined as the part of the diencephalon that develops from the region between the epithalamus and the ventral thalamus. More significantly, it is the part of the diencephalon that has its major efferent connections with telencephalic structures, either striatal or neocortical. In mammals, the neocortical connections dominate, and all dorsal thalamic nuclei project to neocortex. Connections to the striatum are seen for only a few of the nuclei (primarily the intralaminar nuclei) in mammalian brains. All thalamic nuclei have relay cells, which send their axons to the telencephalon, and most have interneurons with locally ramifying axons.

a The Afferents

We have seen that the first order nuclei of the dorsal thalamus receive a significant part of their afferent connections from ascending pathways. Some bring information about the environment to many of the major thalamic nuclei through sensory pathways such as the visual, auditory, somatosensory, or taste pathways. Others bring information about activity in lower, subthalamic centers of the brain such as the cerebellum for the anteroventral and ventrolateral nuclei or the mamillary bodies for the anterior thalamic nuclei (Figures 2 and 3). We shall argue that these can be regarded as the driving inputs for their thalamic nuclei, determining the qualitative characteristics of the receptive fields of the thalamic cells, where these can be defined. Other inputs to first order nuclei are best regarded as modulatory or inhibitory afferents, and these affect quantitative aspects of the receptive field rather than its qualitative structure.⁴ They come from the brain stem, the thalamic reticular nucleus, the hypothalamus, and from the cerebral cortex itself. The thalamic nuclei not outlined in bold in Figure 2 contain higher order circuits and appear to receive most or all of their driving afferents from the cerebral cortex itself, so that the qualitative aspects of their receptive field properties, insofar as they can be defined, depend on cortical, not on ascending, inputs. This distinction is discussed further in later chapters, particularly Chapter VIII. Here it is to be noted that the higher order thalamic relays, in addition to the driving afferents that they receive from cortex, also receive modulatory afferents from cortex and from the other structures noted earlier for the first order nuclei.

The distinction between corticothalamic axons that are drivers and those that are modulators can be made on the basis of the cortical layer from which they arise; current evidence suggests that corticothalamic afferents arising in cortical layer 5 are likely to be drivers, whereas those arising in layer 6 are likely to be modulators (Sherman and Guillery, 1996, 1998; see Chapters III and VIII). In a few instances, discussed in more detail in Chapter VIII, this distinction between drivers and modulators can be demonstrated in functional terms by recording how inactivation of the cortical afferents affects the receptive field properties of thalamic cells, but so far these instances are regrettably rare. Silencing a cortical driver produces a loss of the receptive field, whereas after a modulator is silenced, the receptive field survives. The difference between these two groups of corticothalamic afferents, the drivers and modulators, is seen not only in terms of their origin and their action on receptive field properties of dorsal thalamic cells but also in terms of the structure of the terminals that are formed in the thalamus, and probably also in terms of the nature of the synaptic transfer to thalamic cells. This will be discussed in Chapters III and V.

We shall argue that the thalamus can be regarded as a group of cells concerned, directly or indirectly, with passing information on to the cerebral cortex about essentially everything that is happening in the central or peripheral nervous system. This includes passing